Job Summary
The HPC Middleware Engineer is responsible for deploying, optimizing, and supporting middleware components in a high-performance computing (HPC) environment. This includes scientific libraries, compilers, runtime environments, and container technologies that bridge system software and user applications. The role supports efficient application performance and seamless integration across the HPC stack.
Roles and Responsibilities
- Middleware Stack Management
· Deploy, configure and maintain HPC middleware stack, which includes:
-
- Scientific libraries and frameworks
- HPC runtime environments
· Manage and support container engines for HPC workloads.
· Collaborate with HPC System and Storage Engineers to ensure middleware compatibility with OS, drivers, scheduler, and filesystems.
· Ensure middleware components comply with security policies.
- Application Integration and Support
· Assist users in integrating domain-specific applications with middleware layers to ensure optimal performance and compatibility.
· Assist in tuning and troubleshooting through performance profiling and debugging.
· Provide technical guidance to users on middleware usage and best practices.
- Designing and Planning
· Evaluate new middleware technologies and programming models.
· Contribute to middleware architecture design for future HPC expansions.
Qualifications:
· Degree in Computer Science, Engineering, or related field.
· Minimum 3 years of experience in HPC environments, with focus on middleware or application integration.
· Proficient in Linux systems and CLI tools.
· Strong knowledge of MPI (Cray MPI, OpenMPI, MPICH), scientific libraries (Intel MKL, cuBLAS), and I/O libraries (HDF5, NetCDF).
· Experience with container technologies (Singularity, Enroot, Podman) and HPC programming suites (Intel OneAPI, NVHPC).
· Familiarity with RDMA-based interconnects (InfiniBand, RoCE) and parallel file systems (Lustre, GPFS, BeeGFS).
· Scripting skills in Python, Bash, or Perl.
· Strong analytical and troubleshooting abilities.